What is Goop2?

Goop2 is a peer-to-peer platform for the ephemeral web. Instead of uploading your site to a centralized host, you run a Goop2 peer on your own machine. Your websites, apps, and services exist while you are online and vanish when you shut down -- no servers, no monthly bills, no platform lock-in.

Core ideas

Ephemeral by design -- Content exists only while its owner is present and online. There is no permanent hosting.
Peer-first architecture -- Every participant is both client and server. You serve your own content and browse other peers directly.
No central control -- Discovery happens via local-network mDNS or lightweight rendezvous servers. Nobody owns the network.
Direct connections -- Peers exchange data over libp2p streams. When direct connections aren't possible, a circuit relay helps peers connect through NAT.
Real-time collaboration -- Groups, video calls, file sharing, and cluster compute all work peer-to-peer with no central server in the data path.

How it works

Each peer is a folder on disk containing a configuration file, a site directory, and a local database. When you start a peer, Goop2:

Generates (or loads) a cryptographic identity.
Announces presence on the local network via mDNS and, optionally, to a rendezvous server for WAN discovery.
Serves your site/ directory to any peer that requests it.
Opens a local viewer in your browser so you can manage your site and visit others.

graph TB
    subgraph "Your machine"
        P["Peer"] --> V["Viewer :8080"]
        P --> DB[("data.db")]
        P --> S["site dir"]
    end

    subgraph "Discovery"
        P -->|mDNS| LAN["LAN peers"]
        P -->|WebSocket| RV["Rendezvous server"]
    end

    subgraph "Connectivity"
        RV -->|punch hints| PH["Hole punching"]
        RV -->|relay info| RL["Circuit relay"]
        PH -->|direct| WAN["WAN peers"]
        RL -->|fallback| WAN
    end

Visitors see your site rendered in their own viewer. Data operations (forms, comments, game moves) flow over peer-to-peer streams and are stored in the site owner's local database.

Key concepts

Concept	Description
Peer	One folder + one config + one cryptographic identity.
Site	Static files in the `site/` directory, served to visitors.
Database	A local SQLite database (`data.db`) for dynamic content.
Presence	Soft state: peers announce periodically; absence is inferred after a timeout.
Template	A pre-built application (blog, quiz, game) that bundles HTML, CSS, JS, a database schema, and optional Lua logic.
Group	A real-time communication channel between peers -- used for chat, games, file sharing, and cluster compute.
Rendezvous	A lightweight server that helps peers on different networks find each other. It handles discovery only -- not data.
Relay	A circuit relay (libp2p relay v2) that helps peers behind NAT connect when direct hole-punching fails. Runs alongside the rendezvous server.
MQ	A unified message queue that carries all real-time events (group messages, call signaling, peer announcements) over a single event stream.
Video calls	Peer-to-peer video and audio calls using Pion WebRTC, with WebM streaming for platforms without browser WebRTC support.
Cluster	Distributed compute across peers -- a host dispatches jobs to workers running executor binaries.
File sharing	Upload and share files within a group. Members can browse and download files from any peer in the group.
Bridge	A thin-client mode where peers connect through a WebSocket bridge instead of running a full libp2p node.
Encryption	NaCl key exchange between peers and broadcast key distribution managed by an optional encryption service.

What Goop2 is not

It is not a blockchain or cryptocurrency project.
It is not a CDN or static-site host.
It is not designed for always-on, high-availability services.

Goop2 is best suited for personal sites, small communities, creative experiments, and any context where presence and ephemerality matter more than permanence.